News

Johns Hopkins APL Pioneering Digital Calibration to Create ‘Born-Qualified’ Parts

Steve Storck with the Spectrally Augmented Thermal Understanding Reducing Nonconformance (SATURN) sensor and Process Optimized Laser Adjustments Regulated by Integral Sensors (POLARIS) control system

Steve Storck with the Spectrally Augmented Thermal Understanding Reducing Nonconformance (SATURN) sensor and Process Optimized Laser Adjustments Regulated by Integral Sensors (POLARIS) control system used to augment the performance of the additive manufacturing systems.

Credit: Johns Hopkins APL

Mar 11, 2026
Ajai Raj

The Johns Hopkins Applied Physics Laboratory (APL) in Laurel, Maryland, is developing a comprehensive suite of capabilities to ensure that additively manufactured parts can perform predictably in mission-critical applications — no matter where, when, or on what machines they’re manufactured. This work is poised to revolutionize additive manufacturing (AM) in multiple ways in coming years.

When precisely controlled, AM is capable of producing materials of a high quality for even the most rigorous of defense applications. But currently, AM suffers from a consistency problem.

“Today, AM relies on what I call a ‘guess-and-check’ methodology,” said Steve Storck, chief AM scientist in APL’s Research and Exploratory Development Department. “The engineer doesn’t have data for each new part due to sensor limitations and the complex physics involved, so each build is slightly different. And there are multiple vendors, each with numerous different machines. The problem quickly gets chaotic — it’s a challenge to be sure that parts will survive in a mission scenario.”

As a result, it can take years to design, develop, certify, and begin manufacturing a new part for critical applications like rocket nozzles and aerospace control components. That’s simply not good enough, Storck said.

“We envision a future in which all AM systems produce the same high-quality part based on physical sensor data, independent of the fabrication location and the machine selected for the job,” he said. “To do that, we have to link part quality to fundamental physics.”

Driving Machine Behavior With Thermodynamics

Achieving this level of consistency requires attaining precise control of the heat on the build plates of AM machines — in this case, machines for laser powder bed fusion (LPBF), in which layers of metal powder are fused to create 3D objects with a high-powered laser.

“You could compare the consistency problem with baking a cake in different ovens, in that it’s easy to end up with a product that looks good on the surface but doesn’t function as intended,” Storck explained. “You can end up with a cake that looks delicious but is completely inedible — and you can’t know until you take a bite.”

The APL team realized that the technical challenge is similar to one that’s been solved — achieving consistent image quality on smartphones. Smartphone cameras compensate for variable optical hardware by using software to enhance image quality in real time. Variability in LPBF builds is due to subtle optical discontinuities that affect lasers’ energy input to powders, Storck said.

Correcting for this variability is a major challenge. The optical aberrations involved are incredibly subtle and spatially dependent, requiring measurement and adjustment several orders of magnitude finer than state-of-the-art commercial sensors can provide. This capability seemed so far out of reach that when APL proposed digital calibration using in situ sensing as a solution, some experts in the field said it couldn’t be done. But APL was ready with an answer — an in-house suite of capabilities and facilities designed to address this very problem.

Mary Daffron and Sam Gonzalez inspect a part after fabrication on the EOS M 290 metal 3D printer

Mary Daffron and Sam Gonzalez inspect a part after fabrication on the EOS M 290 metal 3D printer, augmented with advanced control and sensing tools, in the APL Advanced Manufacturing Lab.

Credit: Johns Hopkins APL

JANUS: Predicting Part Performance During Printing

APL has also developed machine learning models that ingest the data collected by SATURN and POLARIS to detect and classify defect-prone print regimes with 95% accuracy, including a novel algorithm that can be used to qualify parts even as they’re being printed. Known as JANUS (Judicious Assessor Noticing Ulterior Signal), the algorithm works similarly to those used for large-scale weather forecasting, except it predicts the formation of defects in additively manufactured parts. The goal is to identify errors and course-correct very early in the printing process — a capability that the defense industrial base has wanted for decades, Storck explained.

“This work is at the leading edge of machine learning, science, and engineering, requiring innovations on all fronts,” said AI research scientist Victor Leon.

JANUS has demonstrated promise to be agnostic to both defect type and part geometry, which can’t be said for other algorithms in the literature, Storck noted. This enables it to be applied to the most challenging real-world use cases, as opposed to cases contrived for research purposes.

“Most of the literature uses trivial cases that don’t apply in real manufacturing contexts, with defects up to ten times the size of the actual part,” he said. “Those models also require millions of parameters, while JANUS only needs 6,000, making it much more lightweight and practical to use.”

SATURN and POLARIS: Controlling Heat and Light

The APL-created Spectrally Augmented Thermal Understanding Reducing Nonconformance (SATURN) instrument is a patented sensor system developed for monitoring the LPBF process.

SATURN features four on-axis photodetectors that are configured to detect specific spectral bands in the visible and infrared wavelengths. At up to 50 MHz, the system measures thermal information at least two orders of magnitude faster than the fastest commercial off-the-shelf solutions.

“The high sensitivity and bandwidth of SATURN allow us to monitor thermal behavior of the melt pool with unprecedented temporal fidelity, while its multispectral detection helps mitigate emissivity effects and potentially confounding optical emissions from plasma,” said Graham Spicer, an optical scientist at APL.

APL researchers are leveraging the multispectral sensing capabilities of the SATURN instrument to collect data on these variations in laser energy input. This, in turn, will make it possible to create a baseline for calibrating LPBF builds so that the exact same part can be produced in the exact same way on any machine, in any location. APL scientists plan to extend this technique beyond LPBF to other AM processes.

The Laboratory’s researchers are also working to expand data analysis to include the quality of the metal powder used, measurements of the shape-related tolerances for specific parts, and other in-process monitoring advancements.

Complementing SATURN is POLARIS (Process Optimized Laser Adjustments Regulated by Integral Sensors), an APL-developed integrated-circuit device that enables precise control of the laser used for LPBF. Initial tests using SATURN and POLARIS have already demonstrated that this level of control and consistency is achievable.

Drawing on Lab-Wide Expertise

Storck emphasized that this work would not have been possible without the combined expertise of a large, multidisciplinary team spanning multiple sectors at APL.

“The Lab uses sensing in a broad range of applications, many of which are relevant to understanding AM processes,” Storck explained. “For instance, understanding the vapor plumes of rockets moving at high speeds also applies to the vapor plumes that form when processing a metal powder with a laser. This work also involves correlating data across a number of distributed sensors, which is another area APL specializes in.”

Transitioning to Industry

This work began with APL funding but has since received investment from multiple sponsors in the Department of War. Between the sensors, algorithms, software, and calibration techniques developed for this purpose, APL has submitted nine invention disclosures, with one issued patent and five patent applications pending to facilitate licensing, according to Susan Ehrlich, APL’s Tech Transfer technology manager.

“By commercializing these in situ monitoring and control capabilities, we can help get them onto commercial machines, support faster qualification of printed parts, and ultimately enable broader adoption that benefits our sponsors, industry partners, and society at large,” Ehrlich said.

Access APL Innovation
Close up of purple gloves and yellow chip

Tech Transfer

APL technology can address industry needs that benefit our sponsors, the public, and the university and may foster economic development.

Explore APL technologies and patents.


Learn more about Tech Transfer

Areas of Impact

Mission Area

The Applied Physics Laboratory, a not-for-profit division of The Johns Hopkins University, meets critical national challenges through the innovative application of science and technology. For more information, visit www.jhuapl.edu.

Media Contact

Amanda Mantiply
Amanda.Mantiply@jhuapl.edu
240-592-6225